RESEARCH ARTICLE

Angiotensin receptor blockers are associated

with lower mortality than ACE inhibitors in
predialytic stage 5 chronic kidney disease: A
nationwide study of therapy with renin-
angiotensin system blockade
Chih-Ching Lin1,2☯‡, Yu-Te Wu2,3☯‡, Wu-Chang Yang2,4, Min-Juei Tsai2,5, Jia-Sin Liu6, Chi-
a1111111111 Yu Yang1,2, Szu-Yuan Li1,2, Shuo-Ming Ou1,2, Der-Cherng Tarng1,2*, Chih-Cheng Hsu6,7,8*
a1111111111
1 Division of Nephrology, Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan,
a1111111111
2 School of Medicine, National Yang-Ming University, Taipei, Taiwan, 3 Sungho Clinic, Ho-Ho-Ho Health
a1111111111 Manage System, Taoyuan, Taiwan, 4 Division of Nephrology, Department of Medicine, Landseed Hospital,
a1111111111 Taoyuan, Taiwan, 5 Department of Medicine, Chang-Hua Hospital, Ministry of Health and Welfare,
Changhua, Taiwan, 6 Institute of Population Health Sciences, National Health Research Institutes, Zhunan,
Taiwan, 7 Department of Health Services Administration, China Medical University, Taichung, Taiwan,
8 Department of Family Medicine, Min-Sheng General Hospital, Taoyuan, Taiwan

☯ These authors contributed equally to this work.

OPEN ACCESS
‡ These authors are co-first authors on this work.
Citation: Lin C-C, Wu Y-T, Yang W-C, Tsai M-J, Liu * dctarng@vghtpe.gov.tw (DCT); cch@nhri.org.tw (CCH)
J-S, Yang C-Y, et al. (2017) Angiotensin receptor
blockers are associated with lower mortality than
ACE inhibitors in predialytic stage 5 chronic kidney
disease: A nationwide study of therapy with renin-
Abstract
angiotensin system blockade. PLoS ONE 12(12):
Dual renin angiotensin system (RAS) blockade using angiotensin-receptor blockers
e0189126. https://doi.org/10.1371/journal.
pone.0189126 (ARBs) in combination with angiotensin converting enzyme inhibitors (ACEIs) is reported
to improve proteinuria in both diabetic and non-diabetic patients. However, its renoprotec-
Editor: Tatsuo Shimosawa, The University of
Tokyo, JAPAN tive effect and safety remain uncertain in patients with advanced chronic kidney disease
(CKD). From January 1, 2000 through June 30, 2009, we enrolled 14,117 pre-dialytic
Received: July 19, 2017
stage 5 CKD patients with serum creatinine >6mg/dL and hematocrit <28% under the
Accepted: November 20, 2017
treatment with erythropoiesis stimulating agents and RAS blockade. We used Cox propor-
Published: December 7, 2017 tional hazards regression models to estimate the hazard ratios (HRs) against the com-
Copyright: © 2017 Lin et al. This is an open access mencement of long-term dialysis and all-cause mortality for ACEI/ARB users. Over a
article distributed under the terms of the Creative median follow-up of 7 months, 9,867 patients (69.9%) required long-term dialysis and
Commons Attribution License, which permits
2,805 (19.9%) died before progression to end-stage renal disease requiring dialysis. In
unrestricted use, distribution, and reproduction in
any medium, provided the original author and comparison with the ARB-only users, dual blockade with ACEIs and ARBs was associated
source are credited. with a significantly higher risk of (1) death in all CKD patients (HR = 1.49, [95%CI, 1.30–
Data Availability Statement: All relevant data are 1.71]; P = 0.02) and in diabetic subgroup (HR = 1.58, [95%CI, 1.34–1.86]; P = 0.02); (2)
within the paper and its Supporting Information composite endpoint of long-term dialysis or death in diabetic subgroup (HR = 1.10, [95%
files. CI, 1.01–1.20]; P = 0.04); (3) hyperkalemia-associated hospitalization in non-diabetic
Funding: This work was supported by intramural subgroup (HR, 2.74, [95%CI, 1.05–7.15]; P = 0.04). However, ACEIs users were associ-
grants (V101C-188, V102C-060, V102C-129, ated with higher mortality than ARBs users in all CKD patients (HR = 1.17, [95%CI, 1.07–
V103C-043, V104C-026, V104C-050, V105C-075,
1.27]; P = 0.03) and in diabetic subgroup (HR = 1.32, [95%CI, 1.18–1.48]; P = 0.03).
V106C-036) and grants for the Integrated Genome
Project (V102E2-001) from Taipei Veterans General
Hospital, and the National Science Council

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 1 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

(NSC101-2314-B010-024-MY3), Foundation for Monotherapy of RAS blockade, especially ARB, is more effective and safer than dual RAS
Poison Control, and the Ministry of Science and blockade in pre-dialytic stage 5 CKD patients.
Technology (MOST 104-2314-B-010 -032 -MY3)
in Taiwan.

Competing interests: The authors have declared

that no competing interests exist.

Introduction
Angiotensin-converting enzyme inhibitor (ACEI) or angiotensin II receptor blocker (ARB)
has been prescribed worldwide to improve proteinuria and delay the progression of chronic
kidney disease (CKD) in both diabetic and non-diabetic patients. Several investigations have
documented its benefit for renal protection to the patients with early CKD (serum creatinine
level: 1.5–3.0 mg/dl)[1, 2] and non-diabetic stage 4 CKD (glomerular filtration rate:15–29 ml/
min/1.73m2 or serum creatinine level: 3.0–5.0 mg/dl).[3] To explore whether ACEI/ARB ther-
apy is the same effective to those patients with advanced CKD at the pre-dialytic stage, our task
group developed a national-wide retrospective study by including all CKD patients diagnosed
between January 1, 2000 and June 30, 2009 in Taiwan, who had serum creatinine level >6 mg/
dl and hematocrit level <28%, and could receive erythropoiesis-stimulating agent (ESA).
Among 28,497 advanced CKD patients, 14,117 ACEI/ARB users, as compared with non-users,
showed to run a significantly lower risk of long-term dialysis (HR, 0.94 [95% CI, 0.91–0.97])
and the composite outcome of long-term dialysis or death (0.94[0.92–0.97]).[4] Thus, the sur-
vival benefit of ACEI or ARB therapy can persist throughout the whole CKD stage, even in
pre-dialytic patients.
Previous investigations have disclosed that dual renin angiotensin system (RAS) blockade
(combination therapy with an ACEI and an ARB) is more effective in proteinuria reduction,
which may provide additional cardiovascular or renoprotective benefit, than either drug alone
in renal disease.[5] However, in the Ongoing Telmisartan Alone and in Combination with
Ramipril Global Endpoint Trial (ONTARGET), the authors found that combination therapy
with an ACEI and an ARB, compared with monotherapy, did not provide more cardiovascular
or renal benefits but increased risk of hyperkalemia and acute kidney injury in persons run-
ning an increased cardiovascular risk.[6] Another recent meta-analysis for patients with early
CKD (stage 1–3) showed no significant difference, either, between dual ACEI plus ARB com-
bination therapy and monotherapy in reducing mortality risk or delay ESRD development.[7]
However, investigation focusing on the safety and effectiveness of dual RAS blockade in
advanced CKD patients, especially at pre-dialytic stage, is lacking. To bridge the gap in the
transition from CKD to ESRD, we assessed the association of the choice of treatment (dual
RAS blockade vs monotherapy) with the risk of long-term dialysis and/or death in this nation-
wide, large cohort of patients with pre-dialytic stage-5 CKD who had hypertension and ane-
mia, and were treated with ESAs.

Materials and methods

Data source
The present study used data from Taiwan’s National Health Insurance (NHI) Research Data-
base which was launched in 1995, managed and released to the public by the National Health
Research Institute of Taiwan, and up to the present covers more than 99%, approximating 23
million, of the residents in Taiwan. This mandatory universal program offers all their medical
records, including date of birth, sex, diagnostic codes, medical procedure and prescription of
drugs. Diseases are coded according to the 2001 International Classification of Diseases, ninth
revision, Clinical Modification (ICD-9-CM). Any information that would expose the identities

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 2 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

of individual patients is de-identified. Having been taken as the primary source for several
published studies, NHIRD has also had the accuracy of diagnoses be repeatedly validated).
This study was approved by the Institutional Review Board of Taipei Veterans General
Hospital.

Study design
This nationwide, retrospective cohort study was performed in Taiwan to determine the associ-
ation between ACEI/ARB usage and the prognosis of advanced CKD. Patients with a primary
diagnosis of CKD (ICD-9 codes 016.0, 042, 095.4, 189, 223, 236.9, 250.4, 271.4, 274.1, 403–404,
440.1, 442.1, 446.21, 447.3, 572.4, 580–589, 590–591, 593, 642.1, 646.2, 753, and 984) subjected
to ESA treatment combined with RAS blockade agent (ACEI or ARB) from January 1, 2000
through June 30, 2009 were enrolled into this study. The National Health Insurance has set the
level of serum creatinine >6 mg/dl and hematocrit <28% as the indication for the use of ESA.
According to the national report of the Department of Health, 85% of the patients with stage 5
CKD were subjected to ESA treatment if they were not on the list requiring renal replacement
therapy. Since the median hematocrit level reported at the initiation of dialysis was 24.4% in
an interquartile range from 20.6% to 27.5% in Taiwan, the cohort of the subjects selected in
the current study should therefore be the most representative of patients with pre-dialytic
stage 5 CKD. The use of ACEI/ARB was defined as prescription within 90 days after the first
ESA treatment. To prevent immortal date bias, the 91th day after the first ESA treatment was
re-defined as the index date. Patients who did not receive ACEI/ARB, who were younger than
20 years or older than 100 years of age, or who had received renal replacement therapy (includ-
ing dialysis and renal transplantation) before ESA treatment were excluded from this study.
Totally 24,765 patients were included in the current study. The subjects who had received
ACEI only, ARB only, or the combined therapy with ACEI plus ARB within 90 days after the
first ESA prescription were defined as ACEI group, ARB group, or ACEI +ARB group accord-
ingly. The remaining patients who took both ACEI and ARB without combination during this
period were defined as ACEI/ARB group. Analyses were all conducted on an intention-to-
treat basis according to the patients’ initial assignment regardless any subsequent changes in
their ACEI/ARB regimen.

Baseline characteristics
Baseline demographic characteristics, including age, gender, comorbidity, geographic location,
nephrologist visits and medications were extracted and recorded. Nephrologist visits and
comorbidities, including diabetes mellitus, stroke sequela, coronary artery disease and cancer,
were defined as record within 3 years before the index date. Charlson comorbidity index
(CCI) scores were used to determine overall systemic health. Each increase in the CCI score
was associated with a stepwise increase of cumulative mortality. Anti-hypertensive agents
other than ACEI/ARB analyzed in this study included diuretics, α-blockers, β-blockers, and
calcium channel blockers. Medications that may potentially influence potassium balance were
also excluded, e.g. Calcium polystyrene sulfonate, Sodium polystyrene sulfonate and sodium
bicarbonate.

Renal outcome and mortality

The observation period started from the index date till the patients died or began long-term
dialysis, or December 31, 2009, whichever occurred earlier. The renal outcome was defined
as the status when the patients entered long-term dialysis. The composite outcome of long-
term dialysis or death referred to either the starting date of long-term dialysis or death,

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 3 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

whichever came first. The first event of hospitalization associated with a diagnosis of hyper-
kalemia (ICD-9 code 276.7) during observation period was defined as hyperkalemia-associ-
ated hospitalization.

Statistical analysis
Baseline characteristics were compared by one-way ANOVA for continuous variables and
Chi-Square test for categorical variables. In the multivariate Cox proportional hazards regres-
sion models, the effects of ACEI, ARB, ACEI+ARB, or ACEIs/ARBs were further adjusted for
age, sex, Charlson comorbidity index (CCI), diabetes mellitus, coronary artery disease, stroke,
cancer, frequency of visits to nephrologists within 3 years before the index date (0, 1–6,or >6
visits), geographic location (northern, middle, southern, or eastern/other islands, according to
NHI registration location), and types of non-ACEI/non-ARB anti-hypertensive agents used.
Study entry was defined as the index date. Results were expressed as hazard ratios (HRs) com-
pared with ARB user. The proportional hazards assumption, the constant HR over time, was
evaluated by comparing the estimated log-log survival curves for all time-independent covari-
ates. All the log-log survival plots assessed graphically showed 2 parallel lines, indicating no
violation of the assumption. Adjusted HRs for long-term dialysis and the composite outcome
associated with ACEI/ARB use were further analyzed among subgroups based on participants’
characteristics (see below). The cumulative hazards of long-term dialysis and the composite
outcome over time were compared among the ACEI, ARB, ACEI/ARB and ACEI + ARB
groups by using the Nelson-Aalen method to adjust covariates adopted in the Cox propor-
tional hazards regression models. All P values were 2-sided, and the level of significance was
set at .05. Analyses were performed using commercially available software (SAS, version 9.2
[SAS Institute Inc.], and Stata SE, version 11.0 [Stata Corp]).

Results
Characteristics of the study population
We identified 24,765 pre-dialytic advanced CKD patients between January 1, 2000 and January
31, 2009 who met the inclusion criteria. Among them, 10,648 patients died or commenced
chronic dialysis within 90 days after the index date. The remaining 14,117 patients were
enrolled into the study and classified into four groups as ARB only (N = 8,203, 58.1%), ACEI
only (N = 3,810, 26.9%), ACEI/ARB (N = 1,095, 7.7%) and concurrent ACEI+ARB use
(N = 1,009, 7.1%) according to their drug prescriptions (Fig 1). The mean age of each group
was 64.5, 65.0, 64.2, and 65.0 years, respectively. The distribution of these groups was slightly
female (53.4%) and elderly (> = 65 years old; 54.3%) predominant. Compared with ARB user,
ACEI user had similar comorbid conditions and medication prescription. However, concur-
rent ACEI+ARB users, as compared with ARB users, were less comorbid with Diabetes (29.3%
vs 58.8%), but more with cardiovascular comorbidities such as myocardial infarction, conges-
tive heart failure and stroke. Concurrent ACEI+ARB users also took more other antihyperten-
sive agents such as alpha-blocker, beta-blocker, calcium channel blocker and diuretics, sodium
bicarbonate and potassium lowering agent, compared with users of either ARB or ACEI.
(Table 1)

Renoprotective effects of ACEIs/ARBs in patients with predialysis

advanced CKD
In a median follow-up of 7 months, 9,867 patients (69.9%) required long-term dialysis and
2,805 (19.9%) died before progression to end-stage renal disease requiring dialysis. The

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 4 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

Fig 1. Flowchart of patient enrollment.

https://doi.org/10.1371/journal.pone.0189126.g001

incidence rates of long-term dialysis, death or the composite outcome of dialysis or death are
listed in Table 2. The adjusted cumulative hazards of long-term dialysis (Fig 2A) and dialysis
or death (Fig 2B) were illustrated as Nelson-Aalen curves for each treatment groups compared
with ARB treatment group. The risk of long-term dialysis or the composite outcome of long-

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 5 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

Table 1. Baseline characteristics of pre-dialysis stage 5 CKD patients, by treatment options.

Treatment ARB only ACEI only ACEI /ARB ACEI and ARB P value
N = 8,203 N = 3,810 N = 1,095 N = 1,009
Age, mean (SD), y 64.5 (12.9) 65.0(13.3) 64.2 (13.1) 65.0 (13.5)
Age, group, y 0.03
20–44 583 (7.1) 288 (7.6) 73 (6.7) 88 (8.7)
45–64 3,219 (39.2) 1,404 (36.9) 436 (39.8) 361 (35.8)
65–74 2,423 (29.5) 1,105 (29) 326 (29.8) 297 (29.4)
75–100 1,978 (24.1) 1,013 (26.6) 260 (23.7) 263 (26.1)
Gender 0.17
Male 3,786 (46.2) 1,832 (48.1) 495 (45.2) 464 (46)
Comorbid conditions within 3 y before the index date
Diabetes 4,826 (58.8) 2,000 (52.5) 315 (28.8) 296 (29.3) <0.01
MI 2,100 (25.6) 969 (25.4) 673 (61.5) 649 (64.3) <0.01
CHF 1,117 (13.6) 547 (14.4) 178 (16.3) 179 (17.7) <0.01
AF 162 (2) 85 (2.2) 24 (2.2) 20 (2) 0.81
Stroke 1,536 (18.7) 704 (18.5) 217 (19.8) 253 (25.1) <0.01
PAOD 108 (1.3) 54 (1.4) 14 (1.3) 20 (2) 0.39
Cancer 658 (8) 353 (9.3) 91 (8.3) 70 (6.9) 0.05
Charlson Comorbidity Index score <0.01
<3 2,887 (35.2) 1,451 (38.1) 358 (32.7) 314 (31.1)
4–5 3,143 (38.3) 1,246 (32.7) 385 (35.2) 363 (36)
>5 2,173 (26.5) 1,113 (29.2) 352 (32.2) 332 (32.9)
Mean(SD) 4.4 (2.2) 4.4 (2.4) 4.6 (2.3) 4.7 (2.3)
Nephrologist visits within 3 y before the index date <0.01
0 1,400 (17.1) 987 (25.9) 271 (24.8) 212 (21)
1–6 1,958 (23.9) 1,093 (28.7) 365 (33.3) 217 (21.5)
>6 4,845 (59.1) 1,730 (45.4) 459 (41.9) 580 (57.5)
Anti-hypertensive drugs used
Alpha-blockers 1,824 (22.2) 897 (23.5) 284 (25.9) 322 (31.9) <0.01
Beta-blockers 3,489 (42.5) 1,478 (38.8) 566 (51.7) 520 (51.5) <0.01
Calcium channel blockers
Non-DHP 780 (9.5) 464 (12.2) 166 (15.2) 264 (26.2) <0.01
DHP 5,768 (70.3) 2,509 (65.9) 829 (75.7) 778 (77.1) <0.01
Diuretics
Thiazides 804 (9.8) 356 (9.3) 143 (13.1) 137 (13.6) <0.01
Loop diuretics 4,967 (60.6) 2,219 (58.2) 757 (69.1) 668 (66.2) <0.01
Other Anti-hypertensive drugs 862 (10.5) 443 (11.6) 202 (18.5) 233 (23.1) <0.01
NaHCO3 1,249 (15.2) 675 (17.7) 202 (18.5) 267 (26.5) <0.01
Calcium polystyrene sulfonate or sodium polystyrene sulfonate 1,622 (19.8) 717 (18.8) 213 (19.5) 345 (34.2) <0.01
Geographic location in Taiwan <0.01
Northern 3,627 (44.2) 1,391 (36.5) 300 (27.4) 343 (34)
Middle 1,820 (22.2) 1,048 (27.5) 332 (30.3) 175 (17.3)
Southern 2,521 (30.7) 1,301 (34.2) 436 (39.8) 479 (47.5)
Eastern or other islands 235 (2.9) 70 (1.8) 27 (2.5) 12 (1.2)

https://doi.org/10.1371/journal.pone.0189126.t001

term dialysis or death was not significantly different between any two groups. In comparison
with ARB only group, however, the risk of death was significantly higher in both ACEI only
group (aHR = 1.17, [95%CI, 1.07–1.27]; P = 0.03) and dual blockade group with concurrent
use of ACEI and ARB (aHR = 1.49, [95%CI, 1.30–1.71]; P = 0.02)

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 6 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

Table 2. Risk of chronic dialysis, dialysis or death, and hyperkalemia associated hospitalization in pre-dialysis stage 5 CKD subjects using ACEI/
ARB treatment.
Subjects Incidence rate (per 100 person-years) Adjusted HR (95% CI)
ACEI/ARB Dialysis Death Dialysis or Hyper- Dialysis Death Dialysis or Hyperkalemia
death kalemia death
All
ARB only 8,203 72.5 17.6 90.1 0.42 1.0 (ref.) 1.0 (ref.) 1.0 (ref.) 1.0 (ref.)
ACEI only 3,810 63.8 21.5 85.3 0.30 0.96 (0.92– 1.17 (1.07–1.27) * 1.01(0.96– 0.76 (0.52–1.10)
1.01) 1.04)
ACEI /ARB 1,095 75.9 21.6 97.5 0.47 1.02 (0.94– 1.10 (0.95–1.27) 1.03 (0.97– 0.96 (0.57–1.64)
1.10) 1.10)
ACEI and 1,009 71.1 32.4 103.5 0.68 0.95 (0.87– 1.49 (1.30–1.71) ** 1.07 (0.99– 1.41 (0.84–2.35)
ARB 1.04) 1.15)
With DM
ARB only 4,826 84.2 20.4 104.6 0.63 1.0 (ref.) 1.0 (ref.) 1.0 (ref.) 1.0 (ref.)
ACEI only 2,000 73.1 29.5 102.6 0.46 0.96 (0.90– 1.32 (1.18–1.48) * 1.03 (0.98– 0.74 (0.47–1.14)
1.02) 1.09)
ACEI /ARB 673 86.8 26.8 113.6 0.54 1.02 (0.93– 1.17 (0.98–1.40) 1.05 (0.96– 0.76 (0.39–1.48)
1.13) 1.14)
ACEI and 649 81.4 40.2 121.6 0.81 0.96 (0.86– 1.58 (1.34–1.86) ** 1.10 (1.01– 1.13 (0.61–2.09)
ARB 1.07) 1.2) *
Without DM
ARB only 3,377 60.9 14.8 75.7 0.19 1.0 (ref.) 1.0 (ref.) 1.0 (ref.) 1.0 (ref.)
ACEI only 1,810 56.7 15.6 72.3 0.18 0.96 (0.90– 0.99 (0.86–1.13) 0.97 (0.91– 0.79 (0.39–1.58)
1.03) 1.03)
ACEI /ARB 422 63.9 15.9 79.8 0.38 1.01 (0.90– 1.02 (0.80–1.30) 1.01 (0.91– 1.77 (0.70–4.44)
1.14) 1.13)
ACEI and 360 47.9 18.7 66.6 0.51 0.95 (0.82– 1.46 (1.16–1.85) * 1.06 (0.94– 2.74 (1.05–
ARB 1.09) 1.19) 7.15) +
+
p value = 0.04,
*p value = 0.03,
** p value = 0.02,
IR: incidence rate, per 100 person-years.
Multivariate analysis was adjusted for variables as listed in Table 1, ref.: reference

https://doi.org/10.1371/journal.pone.0189126.t002

When patients were stratified into diabetic and non-diabetic subgroups by taking the status
of the patients treated with ARB only as reference index, dual blockade with concurrent use of
ACEI and ARB was associated with a significantly higher risk of the composite outcome of
long-term dialysis or death only in diabetic subgroup (Fig 2C, aHR = 1.10, [95%CI, 1.01–1.20];
P = 0.03), but not in non-diabetic subgroup (aHR = 1.06, [95%CI, 0.94–1.19]). In the diabetic
subgroup, the risk of death was significantly higher in both ACEI only users (aHR = 1.32, [95%
CI, 1.18–1.48]; P = 0.03) and users of dual blockade with ACEI and ARB (aHR = 1.58, [95%CI,
1.34–1.86]; P = 0.02). However, in the non-diabetic subgroup, the risk of death was signifi-
cantly higher in users of dual blockade with ACEI and ARB (aHR = 1.46, [95%CI, 1.16–1.85];
P = 0.03) but not in ACEI only users (aHR = 0.99, [95%CI, 0.86–1.13]). In terms of the risk of
long-term dialysis, there was no significantly difference between any groups of different RAS
blockade users in either diabetic or non-diabetic patients.
When patients were stratified by time period as pre-dialysis and under dialysis subgroups
as shown in S1 and S2 Tables. In comparison with ARB only users, ACEI only users and users
of dual blockade with ACEI and ARB remained associated with a significantly higher risk of
death after dialysis in all patients (aHR = 1.61; 1.83 vs. 1), in DM subgroup (aHR = 1.75; 1.76

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 7 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

Fig 2. The adjusted cumulative hazards of clinical outcomes. A: The adjusted cumulative hazards of long-term dialysis. B: The adjusted
cumulative hazards of long-term dialysis or death. C: The adjusted cumulative hazards of long-term dialysis or death in diabetic sub-group.
D: The adjusted cumulative hazards of hyperkalemia associated hospitalization. E: The adjusted cumulative hazards of hyperkalemia
associated hospitalization in non-diabetic subgroup.
https://doi.org/10.1371/journal.pone.0189126.g002

vs.1) as well as in non-DM subgroup (aHR = 1.41; 2.18 vs. 1) (S2 Table). In addition, as com-
pared with ARB-only users, the risk of death (aHR) after dialysis was higher than that before
dialysis for ACEI-only users, not only in all patients (1.61 vs. 1.17) but also in non-DM sub-
group (1.41 vs. 0.99) as well as DM subgroup (1.75 vs. 1.31).

Hyperkalemia-related hospitalization
Compared with ARB users, concurrent ACEI+ARB users showed to run a higher, but not to a
significant level, risk of hyperkalemia-associated hospitalization (aHR of 1.41; 95% CI, 0.84–
2.35). (Table 2, Fig 2D) However, when we stratified patients according to their status of diabe-
tes, dual blockade was associated with a significantly higher risk of hyperkalemia- associated
hospitalization in non-diabetic CKD patients (HR, 2.74, [95%CI, 1.05–7.15]; P = 0.04, Fig 2E)
but not in diabetic patients (HR = 1.13, [95%CI, 0.61–2.09]).

Discussion
Our study demonstrated that dual blockade therapy with ACEIs and ARBs in predialysis CKD
patients was not significantly associated with long-term dialysis or death. Dual RAS blockade
has been thought to be more renoprotective than monotherapy because of its effect on protein-
uria reduction in the short-term treatment.[8] However, some large clinical trials such as
ONTARGET[6] have otherwise disclosed that dual RAS blockade, instead of being effective,
might be harmful because of its potential to increase the risk of acute kidney injury, cardiovas-
cular events and hyperkalemia. They thus excluded CKD patients with serum creatinine
above 3 mg/dl from the study. Another randomized control trial, the Veterans Affairs
Nephropathy in Diabetes (VA NEPHRON-D) study, also showed no significant survival bene-
fit but increased risk of hyperkalemia and acute kidney injury with combination therapy than
with monotherapy by enrolling 1,448 diabetic nephropathy patient with CKD stage 1–3, an

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 8 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

estimated glomerular filtration rate (GFR) of 30.0 to 89.9 ml per minute per 1.73 m2.[9] To
our knowledge, there has been no investigation published to evidence the superior safety and
effectiveness of dual RAS blockade to monotherapy in advanced CKD patients. The current
study might fill this gap.
In this study, the subgroup analysis disclosed a 10% higher risk of long-term dialysis or
death among diabetic patients. The different effect of dual RAS blockade and monotherapy
between diabetic and non-diabetic patient has also been reported by Kim et al. in their small
population crossover trial.[10] In their trial, 24 diabetic nephropathy patients and 19 non-dia-
betic renal diseases (IgA nephropathy) patients received a crossover treatment with ramipril
and dual RAS blockade (Ramipril and Candesartan) in 12-week period. Compared with
monotherapy, this intervention seemed to significantly reduce the 24-hour urinary protein
excretion rate in non-diabetic renal disease patients, but made little difference in diabetic
patients. Another prospective crossover trial enrolling 14 patients with IgA nephropathy and
18 with type 2 diabetic CKD conducted by Song et al. revealed that the urinary TGF-beta1
excretion reduced with dual RAS blockade, compared with monotherapy, only in IgA
nephropathy patients,[11] not in the diabetic ones. The possible explanations for this differ-
ence include the pathophysiological mechanism of diabetic nephropathy, which is more het-
erogeneous and complex than non-diabetic renal disease, and several other factors such as
oxidative stress, glycation end-products, and micro-inflammatory response other than intra-
glomerular hypertension and RAS activation.[12]
Few studies were focused on the effect of dual RAS blockade to non-diabetic CKD patient
group. A former Clinical trial on Aliskiren in Type 2 Diabetes and Cardiorenal End Points
(ALTITUDE) choosing the direct renin inhibitor Aliskiren, instead of ACEI, as the second
RAS blockade added on ARB in diabetic patients was terminated early because of increasing
adverse events including stroke, hyperkalemia and renal complications.[13] On the other
hand, another investigation in similar study design using dual blockade combining direct
renin inhibitor plus ARB, compared with ARB monotherapy, benefited the non-diabetic CKD
patient significantly by reducing their proteinuria and slowing renal function decline.[14] In
our study, although more comorbidities, including myocardial infarction, congestive heart
failure, and stroke, were observed in ACEI plus ARB group than in monotherapy group
(Table 1), the outcome of death or long-term dialysis did not make significant difference to
non-diabetic CKD patients. It is worthy to perform further investigation to identify if non-
diabetic CKD patients could tolerate and benefit from dual RAS blockade more than from
monotherapy.
Our finding of increased risk of death not only in dual blockade users but also in ACEI
users as compared with ARB users was supported by the report of Chan et al. who compared
the relative effectiveness of newly add-on ACEIs, ARBs and dual blockades in reducing cardio-
vascular mortality in about 9,300 chronic hemodialysis (HD) patients over 6 years.[15] After
adjustment for risk factors, increased risk of cardiovascular death was observed in 6,866
patients on ACEI with non-ARB antihypertensive agent (HR of 1.27) as well as in 701 patients
initiated on combined ACEI and ARB therapy (HR of 1.45) as compared with 1,758 patients
initiated on an ARB with non-ACEI antihypertensive therapy.
The reason why ACEIs may be associated with higher risk of death could be explained by
the following mechanisms. First, the anti-inflammatory effect of ACEIs may be less potent
than ARBs according to the report by Gamboa et al.[16], who conducted a randomized, dou-
ble-blind 3×3 crossover study of 15 HD patients assigned to placebo, ACEI with ramipril (5
mg/d), and ARB with valsartan (160 mg/d) for 7 days. Although valsartan and ramipril both
lowered interleukin-6 (IL-6) levels during dialysis, ramipril increased IL-1β concentrations
and decreased IL-10 concentrations compared with placebo, leading to the conclusion that

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 9 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

ARB may induce a greater anti-inflammatory effect than ACEI. Second, ACEIs may aggravate
endothelial dysfunction by increasing the plasma levels of asymmetric dimethylarginine
(ADMA) in HD patients, as it was demonstrated in a randomized cross-over study by Gamboa
et al.[17] They found that ADMA levels were significantly increased throughout the dialysis
session during the treatment by ramipril as compared with valsartan or placebo. Furthermore,
they also showed that ACE inhibition increased bradykinin (BK) levels in ESRD patients dur-
ing HD. The mechanism could be explained by the in vitro study that showed the incubation
with BK increased intracellular ADMA concentration through the stimulation of BK B2-recep-
tor in A549 cells.[17]
The inflammatory status in the period after dialysis should be more severe than that before
dialysis because of less excretion of pro-inflammatory cytokines due to lower residual renal
function. Furthermore, DM patients are also associated with higher inflammatory status than
non-DM patients. Thus, the weaker anti-inflammatory effect by ACEI (vs. ARB) may be asso-
ciated with a higher risk of death in either study period (after vs. before dialysis) or patient sub-
group (DM vs. non-DM) with more severe inflammation.
Hyperkalemia was one of the most concerned adverse effects of RAS blockade, either ACEIs
or ARBs. The combination therapy with ACEI plus ARB exhibits even more complete blockade
of RAS capable to elevate serum potassium level to a much higher level than ACEI or ARB
monotherapy can in CKD patients.[18] A previous retrospective analysis of 245,808 patients
have reported that RAS blockade increased the risk of hyperkalemia more in patients with
CKD than in those without CKD (7.67 vs. 2.30 per 100 patient months, p<0.0001). However,
the odds ratio of death from a moderate (K+ 5.5 and < 6.0mg/dl) to severe (K+ 6.0mg/dl)
hyperkalemia event was significantly higher in non-CKD patient than in CKD patients.[19]
CKD patients, compared with normal population, could tolerate hyperkalemia more and
might not exhibit apparent electrocardiographic or cardiovascular manifestations. Jung Nam
An et al. has reported an observational cohort study for 923 patients, including 70.2% CKD
and 40.6% DM patients who were hospitalized due to severe hyperkalemia, to observe a lower
mortality rate in patients with diabetes and CKD. Furthermore, as CKD progressed to higher
stages, the OR of in-hospital mortality decreased.[20] This may help explain why the risk of
hyperkalemia-associated hospitalization increased so significantly with combination therapy
than with monotherapy in our non-diabetic CKD subgroup, but not in diabetic CKD patients.
Distinct from non-diabetic CKD, diabetic nephropathy usually exhibits chronic hyperkalemia
as a result of hyporeninemic hypoaldosteronism, which suppresses potassium secretion from
renal tubule.[21] Dual RAS blockade may not precipitate hyperkalemia in these patients so
strong as in those CKD patients without hyporeninemic hypoaldosteronism syndrome.
The major strengths of our study are the large sample size and its nationally representative
nature, which rely on a comprehensive medical utilization claim data base to include all the
CKD patients as possible. Bias on the date of immortality was eliminated as best as can by
selecting patients who survived beyond 90 days after ESA prescription and could be observed
thereafter.[22] However, several limitations of this study should be addressed. First, the diag-
nosis of advanced CKD came from the administrative claims data and might not be made as
accurate as by clinically standard practice. For example, patients reported with acute kidney
injury on top of chronic kidney injury may exhibit transient elevation of creatinine level to >6
mg/dl. Thus, we restricted the analyses to the patients receiving ESA therapy for at least 2
ambulatory care visits to minimize the bias. Second, the onset of pre-dialytic stage 5 CKD in
this study was defined as the first day of ESA prescription. For those patients who sought med-
ical assistance late or sought alternative treatment, the timing of enrollment may vary and
introduce bias. Third, blood pressure level was not available from database and we analyzed
different kinds of antihypertensive agents use in each group. Although more antihypertensive

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 10 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

agents prescribed in dual RAS blockade users may introduce bias due to poor blood pressure
control, ARB only users, who showed no difference in other antihypertensive agents use from
ACEI only users, still associated with lower mortality. Fourth, the information on several
potential confounders of mortality and hyperkalemia episodes, such as base-line creatinine
and potassium level, severity of proteinuria, body-mass index and so on, were not available
from the database. Finally, the present study enrolled pre-dialytic stage 5 CKD patients who
were receiving ESA treatment. These results may not be extrapolated to all patients with pre-
dialytic CKD.

Conclusion
Our study demonstrates the difference in the effectiveness and safety between ACEIs, ARBs
and dual RAS blockade therapy (ACEI plus ARB) in pre-dialytic stage 5 CKD patients. As
compared with ARBs, dual blockade with ACEIs and ARBs shows a higher risk of death and
composite endpoint of long-term dialysis or death in diabetic subgroup as well as a higher rate
of hyperkalemia-associated hospitalization in non-diabetic subgroup, and ACEIs are associ-
ated with a higher risk of death, especially in diabetic subgroup. Thus, monotherapy of RAS
blockade, especially ARB, is more effective and safer than dual RAS blockade in pre-dialytic
stage 5 CKD patients.

Supporting information
S1 Table. Risk of death before dialysis initiation in pre-dialysis stage 5 CKD subjects using
ACEI/ARB treatment.
(DOCX)
S2 Table. Risk of death after dialysis initiation in pre-dialysis stage 5 CKD subjects using
ACEI/ARB treatment.
(DOCX)
S3 Table. Diagnostic codes for chronic kidney disease and comorbid conditions.
(DOCX)

Acknowledgments
We are grateful to all research staffs participated in this study.

Author Contributions
Conceptualization: Chih-Ching Lin.
Data curation: Chih-Ching Lin, Yu-Te Wu.
Formal analysis: Chih-Ching Lin, Yu-Te Wu, Wu-Chang Yang, Min-Juei Tsai, Jia-Sin Liu,
Chi-Yu Yang, Szu-Yuan Li, Shuo-Ming Ou.
Methodology: Der-Cherng Tarng, Chih-Cheng Hsu.
Resources: Der-Cherng Tarng, Chih-Cheng Hsu.
Writing – original draft: Chih-Ching Lin, Yu-Te Wu.
Writing – review & editing: Chih-Ching Lin, Yu-Te Wu, Wu-Chang Yang, Min-Juei Tsai,
Jia-Sin Liu, Chi-Yu Yang, Szu-Yuan Li, Shuo-Ming Ou, Der-Cherng Tarng, Chih-Cheng
Hsu.

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 11 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

References
1. Brenner BM, Cooper ME, de Zeeuw D, Keane WF, Mitch WE, Parving HH, et al. Effects of losartan on
renal and cardiovascular outcomes in patients with type 2 diabetes and nephropathy. N Engl J Med.
2001; 345(12):861–9. Epub 2001/09/22. https://doi.org/10.1056/NEJMoa011161 PMID: 11565518.
2. Lewis EJ, Hunsicker LG, Clarke WR, Berl T, Pohl MA, Lewis JB, et al. Renoprotective effect of the
angiotensin-receptor antagonist irbesartan in patients with nephropathy due to type 2 diabetes. N Engl
J Med. 2001; 345(12):851–60. Epub 2001/09/22. https://doi.org/10.1056/NEJMoa011303 PMID:
11565517.
3. Hou FF, Xie D, Zhang X, Chen PY, Zhang WR, Liang M, et al. Renoprotection of Optimal Antiproteinuric
Doses (ROAD) Study: a randomized controlled study of benazepril and losartan in chronic renal insuffi-
ciency. J Am Soc Nephrol. 2007; 18(6):1889–98. Epub 2007/05/15. https://doi.org/10.1681/ASN.
2006121372 PMID: 17494885.
4. Hsu TW, Liu JS, Hung SC, Kuo KL, Chang YK, Chen YC, et al. Renoprotective effect of renin-angioten-
sin-aldosterone system blockade in patients with predialysis advanced chronic kidney disease, hyper-
tension, and anemia. JAMA Intern Med. 2014; 174(3):347–54. Epub 2013/12/18. https://doi.org/10.
1001/jamainternmed.2013.12700 PMID: 24343093.
5. Kunz R, Friedrich C, Wolbers M, Mann JF. Meta-analysis: effect of monotherapy and combination ther-
apy with inhibitors of the renin angiotensin system on proteinuria in renal disease. Ann Intern Med.
2008; 148(1):30–48. Epub 2007/11/07. PMID: 17984482.
6. Mann JF, Schmieder RE, McQueen M, Dyal L, Schumacher H, Pogue J, et al. Renal outcomes with tel-
misartan, ramipril, or both, in people at high vascular risk (the ONTARGET study): a multicentre, rando-
mised, double-blind, controlled trial. Lancet. 2008; 372(9638):547–53. Epub 2008/08/19. https://doi.org/
10.1016/S0140-6736(08)61236-2 PMID: 18707986.
7. Fink HA, Ishani A, Taylor BC, Greer NL, MacDonald R, Rossini D, et al. Screening for, monitoring, and
treatment of chronic kidney disease stages 1 to 3: a systematic review for the U.S. Preventive Services
Task Force and for an American College of Physicians Clinical Practice Guideline. Ann Intern Med.
2012; 156(8):570–81. Epub 2012/04/18. PMID: 22508734.
8. Taal MW, Brenner BM. Renoprotective benefits of RAS inhibition: from ACEI to angiotensin II antago-
nists. Kidney Int. 2000; 57(5):1803–17. Epub 2000/05/03. https://doi.org/10.1046/j.1523-1755.2000.
00031.x PMID: 10792600.
9. Fried LF, Emanuele N, Zhang JH, Brophy M, Conner TA, Duckworth W, et al. Combined angiotensin
inhibition for the treatment of diabetic nephropathy. N Engl J Med. 2013; 369(20):1892–903. Epub
2013/11/12. https://doi.org/10.1056/NEJMoa1303154 PMID: 24206457.
10. Kim MJ, Song JH, Suh JH, Lee SW, Kim GA. Additive antiproteinuric effect of combination therapy with
ACE inhibitor and angiotensin II receptor antagonist: differential short-term response between IgA
nephropathy and diabetic nephropathy. Yonsei Med J. 2003; 44(3):463–72. Epub 2003/07/02. https://
doi.org/10.3349/ymj.2003.44.3.463 PMID: 12833584.
11. Song JH, Lee SW, Suh JH, Kim ES, Hong SB, Kim KA, et al. The effects of dual blockade of the renin-
angiotensin system on urinary protein and transforming growth factor-beta excretion in 2 groups of
patients with IgA and diabetic nephropathy. Clin Nephrol. 2003; 60(5):318–26. Epub 2003/12/03. PMID:
14640237.
12. Shikata K, Makino H. Microinflammation in the pathogenesis of diabetic nephropathy. J Diabetes Inves-
tig. 2013; 4(2):142–9. Epub 2013/03/18. https://doi.org/10.1111/jdi.12050 PMID: 24843643.
13. Parving HH, Brenner BM, McMurray JJ, de Zeeuw D, Haffner SM, Solomon SD, et al. Cardiorenal end
points in a trial of aliskiren for type 2 diabetes. N Engl J Med. 2012; 367(23):2204–13. https://doi.org/10.
1056/NEJMoa1208799 PMID: 23121378.
14. Li SY, Chen YT, Yang WC, Tarng DC, Lin CC, Yang CY, et al. Effect of add-on direct renin inhibitor alis-
kiren in patients with non-diabetes related chronic kidney disease. BMC Nephrol. 2012; 13:89. Epub
2012/08/25. https://doi.org/10.1186/1471-2369-13-89 PMID: 22917002.
15. Chan KE, Ikizler TA, Gamboa JL, Yu C, Hakim RM, Brown NJ. Combined angiotensin-converting
enzyme inhibition and receptor blockade associate with increased risk of cardiovascular death in hemo-
dialysis patients. Kidney Int. 2011; 80(9):978–85. Epub 2011/07/22. https://doi.org/10.1038/ki.2011.228
PMID: 21775975.
16. Gamboa JL, Pretorius M, Todd-Tzanetos DR, Luther JM, Yu C, Ikizler TA, et al. Comparative effects of
angiotensin-converting enzyme inhibition and angiotensin-receptor blockade on inflammation during
hemodialysis. J Am Soc Nephrol. 2012; 23(2):334–42. Epub 2011/12/14. https://doi.org/10.1681/ASN.
2011030287 PMID: 22158433.
17. Gamboa JL, Pretorius M, Sprinkel KC, Brown NJ, Ikizler TA. Angiotensin converting enzyme inhibition
increases ADMA concentration in patients on maintenance hemodialysis—a randomized cross-over

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 12 / 13

 ARB with lower mortality than ACEI in stage 5 CKD

study. BMC Nephrol. 2015; 16:167. Epub 2015/10/27. https://doi.org/10.1186/s12882-015-0162-x

PMID: 26494370
18. Ruilope LM, Aldigier JC, Ponticelli C, Oddou-Stock P, Botteri F, Mann JF. Safety of the combination of
valsartan and benazepril in patients with chronic renal disease. European Group for the Investigation of
Valsartan in Chronic Renal Disease. J Hypertens. 2000; 18(1):89–95. Epub 2000/03/04. PMID:
10678548.
19. Einhorn LM, Zhan M, Hsu VD, Walker LD, Moen MF, Seliger SL, et al. The frequency of hyperkalemia
and its significance in chronic kidney disease. Arch Intern Med. 2009; 169(12):1156–62. Epub 2009/06/
24. https://doi.org/10.1001/archinternmed.2009.132 PMID: 19546417
20. An JN, Lee JP, Jeon HJ, Kim do H, Oh YK, Kim YS, et al. Severe hyperkalemia requiring hospitalization:
predictors of mortality. Crit Care. 2012; 16(6):R225. Epub 2012/11/23. https://doi.org/10.1186/cc11872
PMID: 23171442
21. Uribarri J, Oh MS, Carroll HJ. Hyperkalemia in diabetes mellitus. J Diabet Complications. 1990; 4(1):3–
7. Epub 1990/01/01. PMID: 2141843.
22. Zhou Z, Rahme E, Abrahamowicz M, Pilote L. Survival bias associated with time-to-treatment initiation
in drug effectiveness evaluation: a comparison of methods. Am J Epidemiol. 2005; 162(10):1016–23.
Epub 2005/09/30. https://doi.org/10.1093/aje/kwi307 PMID: 16192344.

PLOS ONE | https://doi.org/10.1371/journal.pone.0189126 December 7, 2017 13 / 13